Restaurant Labor Cost Variance: Why Your Weekly Schedule Hits 28% While Your P&L Shows 34%

Restaurant labor cost variance is the gap between the labor percentage you scheduled at the start of the week and the labor percentage that lands on your P&L at month-end — typically 4 to 8 points higher. Scheduled labor counts only base wages against forecasted revenue, while booked labor includes overtime premiums, payroll taxes, workers' comp, manager salaries, training hours, PTO accruals, and shift-by-shift demand misses. Closing the gap requires a fully burdened, daypart-level model — not a weekly scheduling app.

If you run a full-service restaurant and your manager hands you a beautiful 28% scheduled labor cost on Monday, you already know how this story ends. By the 15th of the following month, your accountant emails over the P&L and labor is 34%. Nobody stole anything. The schedule wasn't wrong. The math was just incomplete — and the variance is hiding in seven specific places. This guide walks through each one, with real public benchmarks, so you can build a labor model that actually matches what hits your books.

The Headline Numbers: What Public Restaurant Operators Are Reporting

Start with the benchmark. The National Restaurant Association's 2024 industry profitability analysis found that full-service operators reported median salaries and wages (including benefits) of 36.5% of sales, with limited-service at 31.7%. Those are P&L numbers — fully loaded — not scheduling-tool numbers.

Public-company filings tell the same story with more precision:

Chipotle Mexican Grill reported labor costs of 25.1% of total revenue for full-year 2025, up from 24.7% in 2024. Q4 2025 alone ran at 25.5%, attributed to wage inflation partially offset by menu price increases (Chipotle Q4/FY 2025 results, Feb 3, 2026).
Texas Roadhouse disclosed in its FY2024 10-K that labor as a percentage of total sales reached 32.7% — a 118 basis-point increase year-over-year, driven by 5.2% wage inflation and 3.1% growth in hours per store week.

Two well-managed public chains, separated by service model, and there is a 760 basis-point gap between them — entirely explainable by daypart density, average check, and how many bodies it takes to put plates on tables. That gap is the same gap you are seeing inside your own four walls between scheduled and actual.

Takeaway: Anchor your target to your service model first (QSR ~25%, casual 25-30%, full-service 30-35%), then to your specific concept's prime-cost ceiling. A single industry average is useless.

Why Your Scheduling App Lies to You: The Six Hidden Costs

The 7shifts 2025 Restaurant Workforce Report and Restaurant365's labor-control playbook converge on the same finding: scheduled labor reflects base hourly wages times forecasted hours. P&L labor reflects everything that actually happened. Here are the six items that consistently get omitted from the schedule but always show up on the books.

1. The Fully Burdened Wage

According to the Bureau of Labor Statistics' Employer Costs for Employee Compensation release (June 2025), total compensation for private industry workers averaged $45.65 per hour, of which benefits were $13.58 — 29.8% of the total. For restaurants specifically, the all-in burden is typically 20-25% on top of base wages, comprising:

FICA (Social Security + Medicare): 7.65% of gross wages
Federal and state unemployment (FUTA + SUTA): 0.6% to 6% depending on state and experience rating
Workers' compensation: 2% to 5% for restaurant class codes
Health insurance, PTO, meal benefit: 8% to 15%

If your scheduler says you have 1,000 hours at $18/hour ($18,000 in wages), your P&L will see closer to $22,500 once burden lands. That alone is most of your 6-point variance.

2. Overtime Premiums on Already-Burdened Wages

The OnPay restaurant payroll guide and AccountingTools' labor rate variance definition both flag the same trap: a 40-hour employee who picks up two extra hours triggers a 1.5x premium, and that premium is itself subject to FICA. A $25 base rate becomes $37.50 in overtime, and you owe 7.65% on the full $37.50 — not the $25 base.

Across a 60-person restaurant where eight people creep two hours over per week, that is roughly $1,200 in overtime wages plus $90 in employer FICA per week — over $67,000 per year on a line item nobody scheduled.

3. Manager and Salaried Labor

Scheduling apps generally model hourly staff. Your GM at $75,000, two AMs at $58,000, and a corporate kitchen director allocated 25% to the unit are all sitting in your labor line on the P&L. On a $2.5M restaurant, that is roughly $190,000 — or 7.6 points of labor — that never appears in the weekly schedule.

4. Training and Onboarding Hours

Texas Roadhouse explicitly cited 3.1% growth in hours per store week in its 2024 10-K, partly attributed to training. A new server at $15/hour who shadows for three full shifts is 24 hours of paid labor producing zero covers. The Restaurant365 labor-control reports note this "ramp-up tax" is invisible because it is spread across the schedule but never tagged as training-specific spend.

5. PTO and Sick Pay Accruals

State paid-sick-leave mandates (California, New York, Washington, and 15+ others) require accrual at roughly one hour per 30 hours worked. That is a ~3.3% accrued liability that hits the P&L as it is taken — typically clustered in slow weeks when the staff member is not generating revenue, so the labor percentage on that day spikes.

6. Forecasting Misses on Both Sides

The schedule denominator is forecasted revenue. The P&L denominator is actual revenue. If your forecast was $52,000 for the week and you did $46,000 (a routine 12% miss in casual dining), the same labor dollars are now suddenly 6.5% higher as a percentage of sales — and that's before any of the burden above.

Takeaway: Add a "labor burden multiplier" of 1.22-1.28 to every scheduled hour, plus a fixed monthly add-on for salaried management, before comparing against your target percentage.

A Step-By-Step Reconciliation Framework

Here is the exact reconciliation we recommend running every Monday morning, using last week's actual data alongside the schedule that was published Sunday.

Pull scheduled hours by daypart (open, lunch, transition, dinner, close) from your scheduling tool — 7shifts, HotSchedules, Restaurant365, whatever you use.
Pull punched hours by daypart from your POS or timekeeping system. Calculate the punch variance: punched hours minus scheduled hours. A 3-5% positive variance is normal slop. Above that, you have a clock-in or schedule-discipline problem.
Layer overtime detection. Flag every employee who crossed 40 hours. Apply 0.5x premium on the overtime portion (the base hour is already counted; only the premium is the new spend).
Apply burden multiplier. Multiply all wages — regular and overtime — by 1.22 (low-burden state) to 1.28 (high-burden state) to get FICA, SUI, workers' comp, and benefits.
Add salaried allocation. Divide your monthly salaried labor + benefits by 4.33 weeks. Add this fixed number to every weekly labor figure.
Calculate against actual revenue, not forecast. The variance between forecast and actual revenue is its own separate problem — but for labor reconciliation, you must use real top-line.
Tag every percentage point of variance to a cause. "Overtime: +0.8%. Burden: +5.2%. Salaried: +2.1%. Forecast miss: +1.3%. Unexplained: +0.4%." That last bucket is the only one your scheduler can actually fix next week.

Takeaway: If your unexplained bucket is consistently above 1%, you have a real operational problem. If it is below 1%, your labor is exactly where it should be and your target percentage itself is the issue.

The Daypart Density Problem

Most full-service operators schedule by template — "Tuesday dinner = 3 servers, 2 line cooks, 1 dish" — regardless of whether last Tuesday did 90 covers or 240. Restaurant365 reports that operators who track labor cost percentage by daypart instead of by week typically find 200-400 basis points of waste concentrated in two windows: post-lunch (2-4 PM) and pre-close (9-10 PM).

Build a four-line check against your schedule before you publish it:

What were covers in this daypart on this day-of-week the last 4 weeks? (use the median, not the average — averages get distorted by one big private event)
What is sales-per-labor-hour at your target percentage? (If you target 30% labor and your average check is $35, you need ~$15 in sales per labor hour, which translates to roughly 0.4 covers per labor hour for full-service.)
Does the schedule for this daypart hit that ratio?
If not, can you stagger start/end times in 30-minute increments to match the curve?

Takeaway: Stagger schedules in 30-minute increments rather than 4-hour blocks. A single server starting at 5:30 PM instead of 4:00 PM on a Tuesday is $22.50 saved at $15/hour, $720 per month, $8,640 per year — for one schedule change at one restaurant.

Building the Variance Model in Excel: What Goes in Each Tab

An effective restaurant labor variance Excel template needs five tabs, each with a single job:

Tab 1 — Burden Setup: State-specific SUI rate, your workers' comp class code rate, health insurance per-FTE cost, accrued PTO rate. This drives your burden multiplier.
Tab 2 — Schedule Import: Paste your published schedule. Auto-calculate scheduled hours by employee, position, daypart, and cost center.
Tab 3 — Punch Import: Paste actual punches. Auto-flag overtime, missed meal breaks (relevant in California, Oregon, and other premium-pay states), and clock-in variance.
Tab 4 — Reconciliation: The seven-line variance bridge from scheduled % to actual P&L %, broken out by cause.
Tab 5 — Daypart Density: Sales-per-labor-hour by 30-minute increment, with a heat-map flagging any block below your minimum threshold.

If you build this from scratch in Excel, plan on 12-20 hours of work plus several weeks of debugging against your live data. The formulas around overtime detection (especially in California, where daily overtime kicks in after 8 hours and double-time after 12) are where most homegrown sheets break.

Takeaway: If you operate in California, Nevada, or Colorado, daily overtime rules mean a homegrown schedule-only model will systematically understate cost. Either build the state logic in or use a model that already has it.

What This Means for Your P&L Next Quarter

The operators closing the scheduled-to-actual gap fastest are doing three things: (1) shifting from weekly schedule reviews to daily punch reconciliations; (2) publishing a fully burdened target instead of a base-wage target to their managers; and (3) treating salaried and hourly labor as a single integrated budget rather than two separate line items. Chipotle's 40-basis-point year-over-year creep in 2025 looks small until you remember it represents roughly $50 million in absolute dollars across the system — and they have one of the most sophisticated labor operations in the industry. If they are losing 40 bps to wage inflation despite that infrastructure, an independent operator without a comparable variance model is losing 300-500 bps and calling it "rising labor costs."

A proper restaurant labor variance Excel template — with burden math, overtime detection, daypart density analysis, and a seven-line variance bridge built in — turns this from a monthly post-mortem into a Monday-morning operational decision. That is the difference between knowing your labor is 34% and knowing exactly which 60 basis points to attack next week. ModelStack's restaurant operations templates include the full labor variance model, the daypart density sheet, and the state-by-state overtime logic, ready to drop your schedule and punch data into — no formula debugging required.